Manajemen Data Frame

---

Deskripsi

Manajemen data infeksi dan vaksinasi Covid-19 seluruh negara.

Input

1. Infected Case

   Sumber: Center for Systems Science and Engineering (CSSE) at Johns Hopkins University (GitHub)

   URL: https://github.com/CSSEGISandData/COVID-19

   Terdiri dari tiga data set:

   • Confirmed
   • Recovered
   • Death

2. Population dan Kode Negara

   Sumber: Center for Systems Science and Engineering (CSSE) at Johns Hopkins University (GitHub)

   URL: https://github.com/CSSEGISandData/COVID-19

3. Histori Vaksin

   Sumber: Our World in Data (Github)

   URL: https://github.com/owid/covid-19-data

4. Vaccine manufacture

   Sumber: Our World in Data (Github)

   URL: https://github.com/owid/covid-19-data

Output

1. Data historis (total akumulatif harian) kasus infeksi dan vaksinasi
2. Data terkini dari kasus infeksi dan vaksinasi

Tools

1. Package

   • base
   • tidyverse (dplyr, readr, lubridate)

2. Function

   • read_csv
   • select
   • rename
   • arrange
   • group_by
   • mutate
   • summarise
   • left_join
   • pivot_wider
   • pivot_longer

---

library(tidyverse)
library(lubridate)

Infected Case

Sumber data:

• Center for Systems Science and Engineering (CSSE) at Johns Hopkins University (GitHub)
• URL: https://github.com/CSSEGISandData/COVID-19

Input:

• Terdapat tiga data dalam tiga file berbeda (dalam format sama), yaitu
  • confirmed
  • recovered
  • death
• Merupakan total kumulatif harian
• Tanggal/hari berupa kolom (format wide)
• Sebagian besar hanya sampai level negara, tapi di beberapa negara ditampilkan hingga level provinsi/state,

Output:

• Satu tabel gabungan dari tiga data
• Tabel diubah menjadi format longer, di mana tanggal/hari menjadi satu kolom
• Agregasi hingga level negara
• (Karena format ketiga data, dibuat fungsi untuk pre-processing)

infected <- function(url){
  read_csv(url) %>%
  rename_all(make.names) %>%
  pivot_longer(cols = starts_with("X"),
                 names_to = "last_update",
                 values_to = "total") %>%
  mutate(last_update = mdy(sub("X","", last_update))) %>%
  rename(country = Country.Region) %>%
  group_by(country, last_update) %>%
  summarise(total = sum(total)) %>%
  ungroup()
}

url_confirmed <- "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv"
url_recovered <- "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_recovered_global.csv"
url_death <-     "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv"

confirmed <- read_csv(url_confirmed)
confirmed

# join all together
infected_case <- infected(url_confirmed) %>% 
  left_join(infected(url_recovered), 
            by = c("country", "last_update")) %>%
  left_join(infected(url_death)
            , by = c("country", "last_update")) %>%
  rename(total_confirmed = total.x,
         total_recovered = total.y,
         total_death     = total) %>%
  mutate(total_active    = total_confirmed - (total_recovered + total_death)) %>%
  mutate(across(starts_with("total"), ~replace_na(., 0)))

infected_case

Populasi dan kode negara

Sumber data:

• Center for Systems Science and Engineering (CSSE) at Johns Hopkins University (GitHub)
• URL: https://github.com/CSSEGISandData/COVID-19

Input:

• Kode standar tiap negara dan regional (ISO code) beserta total populasinya

Output:

• Hanya pilih level negara
• Hanya pilih kolom tertentu yang relevan

url <- "https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/UID_ISO_FIPS_LookUp_Table.csv"
country <- read_csv(url)
country

country <- country %>%
  filter(UID == code3) %>%
  select (iso_code = iso3,
          country = Country_Region,
          latitude = Lat,
          longitude = Long_,
          population = Population)

country

Historis Vaksin

Sumber data:

• Our World in Data (Github)
• URL: https://github.com/owid/covid-19-data

Input:

• Total kumulatif harian vaksinasi setiap negara, meliputi
  • total_vaccinations : total vaksinasi
  • people_vaccinated : total penduduk yang sudah divaksin
  • people_fully_vaccinated : total penduduk yang sudah divaksin lengkap
  • total_boosters : total vaksinasi booster
• Adanya missing value

Output:

• Total kumulatif harian vaksinasi setiap negara
• Imput missing value untuk tiap-tiap negara:
  • Jika tanggal sebelumnya tidak ada non-missing, diisi dengan nol
  • Jika tanggal sebelumnya ada non-missing, diisi dengan non-missing terdekat sebelumnya.

url <- "https://raw.githubusercontent.com/owid/covid-19-data/master/public/data/vaccinations/vaccinations.csv"
vaccination <- read_csv(url)
vaccination

vaccination <- vaccination %>%
  select(location:total_boosters) %>%
  mutate(across(total_vaccinations:total_boosters, 
                ~na_if(., 0))) %>%
  group_by(location, iso_code) %>%
  arrange(location, iso_code, date) %>%
  fill(total_vaccinations:total_boosters, 
       .direction = "down") %>%
  ungroup() %>%
  mutate(across(total_vaccinations:total_boosters, 
                ~replace_na(., 0))) %>%
  rename(country = location, last_update = date)

vaccination

Vaccine manufacture

Sumber data:

• Our World in Data (Github)
• URL: https://github.com/owid/covid-19-data

Input/Output:

• Daftar merek vaksin yang digunakan tiap negara
• Merek vaksin disimpan dalam kolom vaccines
• Apabila suatu negara menggunakan lebih dari satu vaksin, maka daftar vaksin dituliskan seluruhnya di kolom tersebut dengan dipisahkan oleh tanda koma ,.
• Tidak ada pemrosesan data

url <- "https://raw.githubusercontent.com/owid/covid-19-data/master/public/data/vaccinations/locations.csv"
vaccine_manufacture <- read_csv(url)
vaccine_manufacture

Data Historis Infected Case VS Vaccination

Input:

• Tiga data frame yang sudah disipakan sebelumnya
  • infected_case
  • country
  • vaccination

Ouput:

• Join infected_case dengan country berdasarkan kolom country (kedua data berasal dari sumber yang sama (JHU), sehingga penulisan nama negara juga sama)
• Join dengan vaccination berdasarkan kolom iso_code (karena data berasal dari sumber berbeda, sangat direkomendasikan untuk melakukan join menggunakan key/code yang sudah baku)
• Hasil akhir berupa total kumulatif harian tiap negara dari infected case dan vaccination

c19_daily <- infected_case %>%
  left_join(country, by = "country") %>%
  select(iso_code, country, population, last_update,
         starts_with("total")) %>%
  left_join(vaccination, by = c("iso_code", "last_update")) %>%
  group_by(iso_code) %>%
  arrange(iso_code, last_update) %>%
  fill(total_vaccinations:total_boosters, 
       .direction = "down") %>%
  ungroup() %>%
  mutate(across(total_vaccinations:total_boosters, 
                ~replace_na(., 0))) %>%
  rename(country = country.x) %>%
  select(-country.y)

c19_daily

c19_daily %>%
  pivot_longer(c(total_confirmed, total_death, total_vaccinations),
                 names_to = "grp",
                 values_to = "total") %>%
  filter(total > 0) %>%
  group_by(iso_code, grp) %>%
  summarise(last_update = min(last_update)) %>%
  ungroup() %>%
  pivot_wider(names_from = grp, values_from = last_update) %>%
  rename(first_confirmed = total_confirmed,
         first_death = total_death,
         first_vaccination = total_vaccinations)

Data Terkini

Input:

• Tiga data frame yang sudah disipakan sebelumnya
  • infected_case
  • country
  • vaccination
  • vaccine_manufacture

Ouput:

• Tabel yang mencerminkan kondisi terakhir
• Total kasus (confirmed, recovered, death, active), vaksinasi serta tanggal terjadi kasus pertama dan vaksin pertama
• Ditambahkan informasi total populasi tiap negara dan merek vaksin yang digunakan

c19_latest <- c19_daily %>%
  # get latest data
  group_by(iso_code, country) %>%
  mutate(latest_date = max(last_update)) %>%
  ungroup() %>%
  filter(latest_date == last_update) %>%
  # get first_event
  left_join(
  c19_daily %>%
    pivot_longer(c(total_confirmed, total_death, total_vaccinations),
                   names_to = "grp",
                   values_to = "total") %>%
    filter(total > 0) %>%
    group_by(iso_code, grp) %>%
    summarise(last_update = min(last_update)) %>%
    ungroup() %>%
    pivot_wider(names_from = grp, values_from = last_update) %>%
    rename(first_confirmed = total_confirmed,
           first_death = total_death,
           first_vaccination = total_vaccinations)) %>%
  # vaccine manufacture
  left_join(vaccine_manufacture, by = "iso_code") %>%
  select(iso_code:last_update, 
         first_confirmed, first_death, 
         total_confirmed:total_active,
         first_vaccination,
         total_vaccinations:total_boosters,
         vaccines_manufacture = vaccines)

c19_latest